In [7]:
forest_fires_df.hist(figsize=(18, 12), bins=40)
plt.show()
import pandas as pd
import numpy as np
import seaborn as sns
import matplotlib.pyplot as plt
import matplotlib
import warnings
warnings.filterwarnings('ignore')
np.random.seed(23)
forest_fires_df = pd.read_csv('forest_fires_dataset.csv')
forest_fires_df
| X | Y | month | day | FFMC | DMC | DC | ISI | temp | RH | wind | rain | area | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 7 | 5 | mar | fri | 86.2 | 26.2 | 94.3 | 5.1 | 8.2 | 51.0 | 6.7 | 0.0 | 0.00 |
| 1 | 7 | 4 | oct | tue | 90.6 | 35.4 | 669.1 | 6.7 | 18.0 | 33.0 | 0.9 | 0.0 | 0.00 |
| 2 | 7 | 4 | oct | sat | 90.6 | 43.7 | 686.9 | 6.7 | 14.6 | 33.0 | 1.3 | 0.0 | 0.00 |
| 3 | 8 | 6 | mar | fri | 91.7 | 33.3 | 77.5 | 9.0 | 8.3 | 97.0 | 4.0 | 0.2 | 0.00 |
| 4 | 8 | 6 | mar | sun | 89.3 | 51.3 | 102.2 | 9.6 | 11.4 | 99.0 | 1.8 | 0.0 | 0.00 |
| ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... |
| 512 | 4 | 3 | aug | sun | 81.6 | 56.7 | 665.6 | 1.9 | 27.8 | 32.0 | 2.7 | 0.0 | 6.44 |
| 513 | 2 | 4 | aug | sun | 81.6 | 56.7 | 665.6 | 1.9 | 21.9 | 71.0 | 5.8 | 0.0 | 54.29 |
| 514 | 7 | 4 | aug | sun | 81.6 | 56.7 | 665.6 | 1.9 | 21.2 | 70.0 | 6.7 | 0.0 | 11.16 |
| 515 | 1 | 4 | aug | sat | 94.4 | 146.0 | 614.7 | 11.3 | 25.6 | 42.0 | 4.0 | 0.0 | 0.00 |
| 516 | 6 | 3 | nov | tue | 79.5 | 3.0 | 106.7 | 1.1 | 11.8 | 31.0 | 4.5 | 0.0 | 0.00 |
517 rows × 13 columns
Rozszyfrujmy niezrozumiałe skróty:
represents fuel moisture of forest litter fuels under the shade of a forest canopy. It is intended to represent moisture conditions for shaded litter fuels, the equivalent of 16-hour timelag. It ranges from 0-101. Subtracting the FFMC value from 100 can provide an estimate for the equivalent (approximately 10h) fuel moisture content, most accurate when FFMC values are roughly above 80.
represents fuel moisture of decomposed organic material underneath the litter. System designers suggest that it is represents moisture conditions for the equivalent of 15-day (or 360 hr) timelag fuels. It is unitless and open ended. It may provide insight to live fuel moisture stress.
much like the Keetch-Byrum Drought Index, represents drying deep into the soil. It approximates moisture conditions for the equivalent of 53-day (1272 hour) timelag fuels. It is unitless, with a maximum value of 1000. Extreme drought conditions have produced DC values near 800.
is analogous to the NFDRS Spread Component (SC). It integrates fuel moisture for fine dead fuels and surface windspeed to estimate a spread potential. ISI is a key input for fire behavior predictions in the FBP system. It is unitless and open ended.
forest_fires_df.info()
<class 'pandas.core.frame.DataFrame'> RangeIndex: 517 entries, 0 to 516 Data columns (total 13 columns): # Column Non-Null Count Dtype --- ------ -------------- ----- 0 X 517 non-null int64 1 Y 517 non-null int64 2 month 517 non-null object 3 day 517 non-null object 4 FFMC 517 non-null float64 5 DMC 517 non-null float64 6 DC 517 non-null float64 7 ISI 517 non-null float64 8 temp 517 non-null float64 9 RH 517 non-null float64 10 wind 517 non-null float64 11 rain 517 non-null float64 12 area 517 non-null float64 dtypes: float64(9), int64(2), object(2) memory usage: 52.6+ KB
forest_fires_df.describe()
| X | Y | FFMC | DMC | DC | ISI | temp | RH | wind | rain | area | |
|---|---|---|---|---|---|---|---|---|---|---|---|
| count | 517.000000 | 517.000000 | 517.000000 | 517.000000 | 517.000000 | 517.000000 | 517.000000 | 517.000000 | 517.000000 | 517.000000 | 517.000000 |
| mean | 4.669246 | 4.299807 | 90.644681 | 110.872340 | 547.940039 | 9.021663 | 18.889168 | 44.288201 | 4.017602 | 0.021663 | 12.847292 |
| std | 2.313778 | 1.229900 | 5.520111 | 64.046482 | 248.066192 | 4.559477 | 5.806625 | 16.317469 | 1.791653 | 0.295959 | 63.655818 |
| min | 1.000000 | 2.000000 | 18.700000 | 1.100000 | 7.900000 | 0.000000 | 2.200000 | 15.000000 | 0.400000 | 0.000000 | 0.000000 |
| 25% | 3.000000 | 4.000000 | 90.200000 | 68.600000 | 437.700000 | 6.500000 | 15.500000 | 33.000000 | 2.700000 | 0.000000 | 0.000000 |
| 50% | 4.000000 | 4.000000 | 91.600000 | 108.300000 | 664.200000 | 8.400000 | 19.300000 | 42.000000 | 4.000000 | 0.000000 | 0.520000 |
| 75% | 7.000000 | 5.000000 | 92.900000 | 142.400000 | 713.900000 | 10.800000 | 22.800000 | 53.000000 | 4.900000 | 0.000000 | 6.570000 |
| max | 9.000000 | 9.000000 | 96.200000 | 291.300000 | 860.600000 | 56.100000 | 33.300000 | 100.000000 | 9.400000 | 6.400000 | 1090.840000 |
Nie ma braków, dane numeryczne i obiekty. describe() automatycznie nie wzięło do analizy kolumn z danymi innymi niż numeryczne, przez co się nie wysypało - fajne!
forest_fires_df.hist(figsize=(18, 12), bins=40)
plt.show()
sns.heatmap(forest_fires_df[['FFMC', 'DMC', 'DC', 'ISI', 'temp', 'RH', 'wind']].corr(), annot=True)
plt.show()
Najwyższa korelacja występuje pomiędzy cechami FFMC i ISI, DMC i DC oraz RH i temperaturą
sns.regplot(data = forest_fires_df, x = 'temp', y = 'RH')
plt.show()
Jak widać wraz ze wzrostem temperatury maleje względna wilgotność.
sns.regplot(data = forest_fires_df, x = 'FFMC', y = 'ISI')
plt.show()
Usuńmy zmienne odstające:
factor = 3
upper_lim1 = forest_fires_df['FFMC'].mean () + forest_fires_df['FFMC'].std () * factor
lower_lim1 = forest_fires_df['FFMC'].mean () - forest_fires_df['FFMC'].std () * factor
data = forest_fires_df[(forest_fires_df['FFMC'] < upper_lim1) & (forest_fires_df['FFMC'] > lower_lim1)]
upper_lim2 = forest_fires_df['ISI'].mean () + forest_fires_df['ISI'].std () * factor
lower_lim2 = forest_fires_df['ISI'].mean () - forest_fires_df['ISI'].std () * factor
data = data[(data['ISI'] < upper_lim2) & (data['ISI'] > lower_lim2)]
sns.regplot(data = data, x = 'FFMC', y = 'ISI')
plt.show()
sns.regplot(data = forest_fires_df, x = 'DC', y = 'DMC')
plt.show()
sns.pairplot(forest_fires_df, y_vars="area", x_vars=forest_fires_df.columns.values[:2], diag_kind=None)
sns.pairplot(forest_fires_df, y_vars="area", x_vars=forest_fires_df.columns.values[4:6], diag_kind=None)
sns.pairplot(forest_fires_df, y_vars="area", x_vars=forest_fires_df.columns.values[6:9], diag_kind=None)
sns.pairplot(forest_fires_df, y_vars="area", x_vars=forest_fires_df.columns.values[9:12], diag_kind=None)
plt.show()
Widzimy kilka zmiennych odstających.
x = (forest_fires_df.groupby(["X", "Y"]).size().reset_index(name='number')).pivot(index="Y", columns="X", values="number").fillna(0)
sns.heatmap(x, cmap = "Reds")
plt.show()
Istnieją obszary, w których pożary występują zdecydowanie najczęściej, oraz takie, w których są one sporadyczne.
month_df = forest_fires_df.loc[:, 'month'].value_counts()
a = pd.DataFrame(month_df)
plt.bar(a.index, a['month'])
plt.show()
day_df = forest_fires_df.loc[:, 'day'].value_counts()
b = pd.DataFrame(day_df)
plt.bar(b.index, b['day'])
plt.show()
Jak widzimy, zdecydowanie najwięcej pożarów miało miejsce w sierpniu i wrześniu. Z kolei jeśli chodzi o dzień tygodnia, przodują dni weekendowe.
from pandas_profiling import ProfileReport
profile = ProfileReport(forest_fires_df, title="Pandas Profiling Report")
profile.to_notebook_iframe()
Narzędzie umożlwia nam wstępną analizę danych. Oczywiście, nie jest ono tak dokładne i "spersonalizowane" jak własnoręczna analiza, ale może być przedatne we wstępnym etapie. Jego sporą wadą jest to, że okazuje się nieoptymalne dla dużych zbiorów danych. Poza tym nie daje nam szczegółowych informacji np. dotyczących rozkładów konkretnych cech.